dissenting. The majority opinion rules that a defendant may be convicted of complicity on the basis of the testimony of more than one accomplice even, though that accomplice’s testimony is uncorroborated by other, non-accomplice testimony. Since I believe that this holding is contrary to the proper interpretation of R. C. 2923.03(D) and to the principles underlying it, I must dissent.
The statute controlling the instant cause is R. C. 2923.-03, which states, in pertinent part:
“(D) No person shall be convicted of complicity under this section solely upon the testimony of an accomplice, unsupported by other evidence.”
The majority determines that a person may be convicted of complicity under R. C. 2923.03(D) on the grounds that the statute does not clearly exclude the uncorroborated evidence of more than one accomplice.5 (In the words of the majority, “the statutory reference is not to ‘testi*114mony of accomplices’ but to ‘the testimony of am accomplice.’ ”) To begin with, the use of the articles “the” and “an” does not clearly limit the effect of R. C. 2923.03(D) to ■one accomplice. “The” is merely a definite article; it does not have any numerical significance. (The General Assembly could just as well have said “the testimony of accomplices,” for instance.) Moreover, the article “an” does not necessarily connote singularity; “an accomplice” may be one accomplice, but it also may be one of many accomplices — that is, a member of a class.6 Given that “an accomplice” is at least as certainly a reference to a class as to one individual and that the General Assembly requires us to strictly construe all statutes defining offenses or penalties in favor of the accused (R. O. 2901.04[A]), I find the majority’s reading of R. 0. 2923.03(D) to be unwarranted.
I also submit that the majority’s holding is illogical and undermines the principles behind R. 0. 2923.03(D). By limiting the effect of accomplice testimony, the General Assembly was recognizing that the testimony of one who stands to gain by pointing the finger at someone else is inherently suspect. In the words of the opinion in State v. Myers (1978), 53 Ohio St. 2d 74, 75, issued this very year:
■ “In this day of plea bargaining and immunized testimony * * * it is vitally important that one implicating an accomplice do something more than point a finger. Hi a testimony must be corroborated by some other fact, circumstance, or testimony which also points to the identity of the one he accuses as a guilty actor.”
Indeed, the requirement of corroboration was based historically on the assumption that the accomplice might expect to purchase immunity from punishment by falsely *115accusing and procuring the conviction of others. 7 Wig-more on Evidence 322 (3 Ed. 1940), Section 2057. Finally, given the human tendency to become more likely to act in inherently dangerous or morally questionable ways when others are doing so as well, it seems to me that the uncorroborated testimony of two accomplices is at least, and maybe even more, suspect than the uncorroborated testimony of one.
The majority argues, however, that “to construe the statute as forbidding a conviction * * * would mean that ‘the testimony of the accomplice amounted to “nothing.” ’ ” It argues further that, because such testimony may be admitted if it is corroborated by non-accomplice testimony, the General Assembly did not intend to treat it as a nullity.
That argument is fallacious. Under R. C. 2923.03(D) the General Assembly treats corroborated and uncorroborated accomplice testimony as entirely separate entities. Therefore, this court cannot infer legislative approval of the latter from its approval of the former. Moreover, R. C. 2923.03(D) treats accomplice testimony which is not corroborated by non-accomplice testimony as a nullity for purposes of conviction. If such testimony by one accomplice is worth nothing, similar testimony by a second accomplice also carries no weight. Zero 4- Zero == Zero.
For the above reasons, I dissent.
Webster’s Third New International Dictionary defines “a,” of which “an” is the variant preceding nouns beginning with vowels, as an “indefinite article * * * used as a function word before most singular nouns * * * when the individual in question is undetermined, unidentified * * *” and when that noun is “an example of [a named class] * *